This is an automated email from the ASF dual-hosted git repository.
snagel pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/nutch.git
The following commit(s) were added to refs/heads/master by this push:
new ddabe9694 NUTCH-3144 URLUtil unit tests fail after upgrade to
crawler-commons 1.6
ddabe9694 is described below
commit ddabe969426e0822d425986fb9c3b1ef7aa56006
Author: Sebastian Nagel <[email protected]>
AuthorDate: Wed Jan 7 23:05:54 2026 +0100
NUTCH-3144 URLUtil unit tests fail after upgrade to crawler-commons 1.6
- adapt unit tests to changes introduced in
https://github.com/crawler-commons/crawler-commons/pull/478
- test for example given in Javadoc of getDomainSuffix
---
src/java/org/apache/nutch/util/URLUtil.java | 2 +-
src/test/org/apache/nutch/util/TestURLUtil.java | 2 ++
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/src/java/org/apache/nutch/util/URLUtil.java
b/src/java/org/apache/nutch/util/URLUtil.java
index afd6f1385..44c6309d2 100644
--- a/src/java/org/apache/nutch/util/URLUtil.java
+++ b/src/java/org/apache/nutch/util/URLUtil.java
@@ -262,7 +262,7 @@ public class URLUtil {
EffectiveTldFinder.EffectiveTLD suffix =
EffectiveTldFinder.getEffectiveTLD(host, true);
if (suffix != null) {
- return suffix.getDomain();
+ return suffix.getSuffix();
}
return null;
diff --git a/src/test/org/apache/nutch/util/TestURLUtil.java
b/src/test/org/apache/nutch/util/TestURLUtil.java
index 32dda0929..092edb9c1 100644
--- a/src/test/org/apache/nutch/util/TestURLUtil.java
+++ b/src/test/org/apache/nutch/util/TestURLUtil.java
@@ -147,6 +147,8 @@ public class TestURLUtil {
assertEquals("2000.hu", URLUtil.getDomainSuffix(url));
// test non-ASCII
+ url = new URL("https://www.taiuru.māori.nz/");
+ assertEquals("xn--mori-qsa.nz", URLUtil.getDomainSuffix(url));
url = new URL("http://www.example.flå.no");
assertEquals("xn--fl-zia.no", URLUtil.getDomainSuffix(url));
url = new URL("http://www.example.栃木.jp");